[SRU][F:linux-bluefield][PATCH 0/2] CT offload fix for crash on stress

Roi Dayan roid at nvidia.com
Sun Apr 11 11:45:21 UTC 2021


BugLink: https://bugs.launchpad.net/bugs/1922672


SRU Justification:

There is a race between netfilter GC updating ct conn timeout and other events
reading the timeout, potentially crashing the kernel.


* brief explanation of fixes

The fix is setting the offload timeout early and not relying on gc.
The fix is already upstream and cherry picked here.

* How to test

Testing was done with stress http traffic opening conns, short data, close conns.
different 5-tuple each time.


* What it could break.

Potentially crash the kernel.



Roi Dayan (2):
  netfilter: conntrack: Move nf_ct_offload_timeout to header file
  netfilter: flowtable: Set offload timeout when adding flow

 include/net/netfilter/nf_conntrack.h | 12 ++++++++++++
 net/netfilter/nf_conntrack_core.c    | 12 ------------
 net/netfilter/nf_flow_table_core.c   |  2 ++
 3 files changed, 14 insertions(+), 12 deletions(-)

-- 
2.26.2




More information about the kernel-team mailing list